A randomized controlled trial assigns 200 patients with high cholesterol to either a statin medication ($n=100$) or placebo ($n=100$). After 6 months:

• Assigned to statin: 80 took medication as prescribed; 20 never started or stopped early
• Assigned to placebo: 95 took placebo; 5 switched to statin

If analysis uses 'intention-to-treat' (ITT)—analyzing participants in their originally assigned groups—what is a potential limitation of this approach?